Text-based analysis to detect figure plagiarism

Abstract

Plagiarism, the process of copying someone else’s text or data without due recognition of the source is a serious academic offence. Many techniques have been proposed for detecting plagiarism in texts but only few techniques exist for detecting figure plagiarism. The main problem associated with existing techniques is that they are not applicable to non-textual elements of figures in research publications. This paper addresses the problem of figure plagiarism in scientific articles and proposes solutions to detect cases where an exact copy or modified figure retains the essential data in the original figure. In this paper, we proposed a deep figure analysis to detect all types of possible figure plagiarism ranging from simple copy and paste to plagiarism of strong modification to the content of the figure source. Unlike existing figure plagiarism detection methods, which compare between figures based on surface features. The proposed method represents each component of a figure and provides information about the text inside its component and the relationships with other component(s) to capture the meaning of the figure. using component-based comparison, and an improvement over existing methods which cannot extract enough information from figures to detect plagiarism. The results obtained by the proposed method are considered as one of the interesting research solutions for figure plagiarism.